The past project periods identified several risk factors for incident AD and examined the biologic pathways linking risk factors to clinically diagnosed disease. We quantified AD pathology, cerebral infarctions, and Lewy bodies in brain tissue from 300 persons and found that one or more of these pathologies is present in virtually everyone with clinically diagnosed AD, most people with mild cognitive impairment (MCI), and many without dementia or MCI. Of note, only the genetic risk factor APOE led to clinical disease directly through an association with AD neuropathology, in addition to a small component through cerebral infarctions. Diabetes appeared to lead to disease through cerebral infarctions. BMI and mild parkinsonian signs, while related to AD pathology, are more likely to be a consequence of pathology and an early sign of clinical disease. Education and conscientiousness were not directly related to AD pathology but modified the relation of pathology to cognition. Depressive symptoms and distress proneness were related to clinical disease through as yet unknown mechanisms. Finally, we did not find a strong relation of statins to clinical or neuropathologic phenotypes. The overall goal of the proposed continuation is to identify novel genetic variants associated with risk of AD. The central idea of the proposal is that the effect of particular alleles on the accumulation of AD neuropathology is more specific and stronger, and therefore more easily detectable, compared to clinically diagnosed AD, which we have demonstrated to be a complex function of AD pathology and many other factors. In Aim 1, we will conduct a genome-wide association scan to identify loci associated with AD neuropathology in deceased participants from the Exploratory Cohort (Religious Orders Study). Aim 2 will conduct a second genome-wide scan in deceased participants from the Confirmatory Cohort (Memory and Aging Project), and in pooled analysis of data from both cohorts, to confirm the loci identified in Aim 1 and identify loci that were missed in Aim 1. Aim 3 will investigate the clinical relevance of the loci identified in Aims 1 and 2 by examining their relation to incident AD, incident MCI, and change in cognitive function over up to 20 years. Aims 4 and 5 will conduct fine mapping and examine gene-gene and gene-environmental interactions on phenotypes relevant to both neuropathologic and clinical phenotypes. Support for the proposal comes from the compelling preliminary data demonstrating the dramatically increased power and much reduced sample size required to identify associations with quantitative neuropathologic phenotypes compared to clinical phenotypes. We are not aware of any other studies of older men and women of comparable size, and follow-up and autopsy rates, in which these analyses can be performed. Thus, the proposed study represents a timely, novel and potentially powerful approach to the identification of genetic variants associated with risk of AD that strongly complements other ongoing large, case-control, family, and population-based epidemiologic studies seeking to identify such variants.